Roozbeh Pournader <roozbeh@Sina.sharif.ac.ir> wrote:
>> I have suggested on this list using Plane 14 tags to assist in glyph
>> selection between C, J, and K or between Russian italics and Serbian
>> italics because I thought they would provide a nice, all-Unicode
>> solution *without* resorting to higher protocols. Other Unicode
>> mechanisms, like LTR and RTL directional overrides and ligation
>> control via ZWJ and ZWNJ (to name only two), seem to have been
>> invented for exactly that purpose.
> I don't agree with the last comment. ZWJ and ZWNJ are not only for
> visual appearance. While the difference between Chinese and Japanese
> glyphs makes no difference in meaning, leaving or using a ZWJ or ZWNJ
> sometimes changes the word meaning. That's at least true for Persian,
> I don't know about Indic languages.
I never said that ZWJ and ZWNJ were useful only for visual rendering
(although I did not know of the case Roozbeh described). However, that
is certainly one of its possible uses. What I did say was that the
ability to control ligature formation (in alphabetic scripts) by means
of ZWJ and ZWNJ was invented so that such ligatures could exist in
plain Unicode text, without the need for external markup.
LTR and RTL directional overrides, which are clearly a visual concept,
similarly exist so that proper directionality can be expressed in plain
Unicode text without markup.
My point regarding language tags was that there is a precedent in
Unicode for expressing some of these "semi-advanced" presentation
concepts -- whose absence might cause incorrect rendering -- in plain
text, and that not everything should be delegated to HTML, XML, or
Tomorrow's Favorite ML.
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:13 EDT